Manage Learn to apply best practices and optimize your operations.

Building a business case for Linux adoption

Making the business case for Linux isn't rocket science, but there's no one magic formula that's going to work for everyone, either.

Making the business case for Linux isn't rocket science, but there's no one magic formula that's going to work for everyone, either. It's got as much to do with knowing your organization -- the key issues and problems and why you're considering adopting open source to begin with -- as it does with systems, upgrades, personnel and other costs.

That said, however, there are a few accepted truisms, according to experts. First, it's much easier to make a case for Linux as part of the server or other IT infrastructure than it is on the desktop. Linux simply isn't as proven on the desktop as it is in other applications, observers note. Second, going to Linux for a brand-new application -- or to replace a big and costly Unix box -- is a much easier and straight-line financial sell than is pitching Linux to replace Windows systems.

"If it's a Windows or Microsoft migration, it's really challenging," said Andrew Aiken, managing partner of the Olliance Group, a consultancy in Palo Alto, Calif., that helps companies implement open source software. The skills needed in a Microsoft shop are not directly transferable to the Linux world, he said, and so there's often a significant amount of money needed for retraining IT professionals and end users -- an amount that many of his clients underestimate, he said. In fact, new (and more expensive) Linux-savvy IT staffers may need to be hired.

Migration and integration costs

Then there are the costs for software migration -- porting or rewriting -- and integration with, say, a new Web browser or back-end customer relationship management package. These costs can add up quickly, especially for an enterprise of any significant size that may have hundreds of applications that touch, or are touched by, any proposed outgoing Microsoft software.

On the other hand, migration from a Unix environment is almost a no-brainer, according to observers. "You're replacing a proprietary RISC box with much less expensive hardware from Intel or AMD," said Bernard Golden, CEO of Navica, a systems integrator in San Carlos, Calif. "You can replace a box that costs $100,000 with one that costs $10,000. There's almost an intuitive sense that it's cheaper."

And because Linux and Unix are so similar, software migration and porting costs are low or nonexistent, as are personnel and training issues.

Hardware savings

This has been the case at the U.S. Naval Oceanographic Office (NAVO), which gathers data with its fleet of six survey ships and provides information to other units within the Navy. The Open Source Software Institute (OSSI) in Oxford, Miss., helped conduct a two-year, cost-benefit study about open source for NAVO. Total hardware savings alone for two Unix-based applications that moved to Linux amounted to almost $1.2 million. One was an image processing application migrated from SGI Irix workstations to notebooks running Linux, and the second was a shipboard data collection system that ran on PA-RISC systems under HP-UX and now run on Intel-based platforms under Linux.

"NAVO was already going with Linux on its own," said John Weathersby, executive director of OSSI. "We just helped them quantify the savings."

ROI analysis

Still, most organizations are like NAVO: They bring in open source and then worry about the business case later, only when they need to, because they've reached some kind of critical mass or are considering Linux for a substantial or bet-the-business application. "It's amazing, but only around 10% of all companies do any kind of thorough" analysis when looking at Linux or open source in general, Aiken said. The reason is that organizations typically start by deploying Linux to run a file server or by putting Apache on the Web server. With each open source success, increasing numbers of open source applications find their way into the company, more in an ad hoc way than in any real planned or formal fashion.

Whenever the time is right for a business case, though, there are other things to consider.

One approach is to consider a five-year timeline, and look at things like support costs, software licensing costs and the need for hardware upgrades that may be required every few years to accommodate whatever new software you'd need to upgrade to, said Gordon Haff, senior analyst at Illuminata, a consultancy in Nashua, N.H.

"A lot of the things going into ROI can be hard to determine numbers for, but you know they're not zero," Haff said. And if you ignore them, you're assigning a value of zero to them." These include soft costs like a feeling of greater control and not being tied to a specific vendor.

Other factors include systems uptime or availability, the time and money spent on patching buggy software, and not being as vulnerable to viruses or worms.

Haff also suggested that "If you're looking at a shift on the desktop, it's absolutely relevant to look at the current cost per desktop per year today, taking into account all costs. That's the baseline. If you get up to what you're paying today, it's probably not a good idea" to pitch a switch. Or, as Navica's Golden said, "Why change for the same functionality?"

Still, Haff said, "Don't let the numbers blind you to the obvious, either. If there are more strategic reasons to go with one platform over another, then an ROI analysis shouldn't stand in the way. There might be business benefits underlying the reason you want to go with open source," but don't just suggest it only because it's open source.

Another thing not to do: Avoid looking at the initial acquisition costs of Linux as the entire cost savings, said Russ Pavlicek, a longtime Linux proponent who works as a senior architect for Cassatt Corp., a systems management vendor. "When people talk about Linux being low-cost, they're almost always talking about the cost of acquisition," he said. "For a startup or small company this may be significant, but it's not the whole picture. In a larger organization, the other costs of running an infrastructure are much more significant."

Benefits of open source

In Pavlicek's opinion, the top reason for going with open source is control. "It lets you balance between being able to go with off-the-shelf software that you can then customize" to make the software really work the way you want it to, he said. "It allows you to close the gap between where you are and where you want to be."

This is assuming the IT organization is up to the task, however. "If you give the same open source code to another IT group that's dull-headed, they will still be as lost as before because they don't know how to use it," Pavlicek said. But in the right hands, open source allows the organization that's sharpest to leverage its IT talent to help the company accomplish new things or do old things like customer service in new ways and at relatively low cost.

Above all, Pavlicek said, "If the metric used to judge anything isn't fair, then it's not going to work. You need to find the measurement for your organization that's most relevant, then make sure it's fair and complete. Apply it and let the chips fall where they may."

Johanna Ambrosio is a freelance writer based out of Marlborough, Mass. She can be reached at

Dig Deeper on Enterprise business applications

Start the conversation

Send me notifications when other members comment.

Please create a username to comment.